The cross-sectional anatomy of the intervertebral disc and spinal canal can be studied in detail by CT whichis not invasive and is much simpler than conventional myelography. The shape of the normal disc and distributionof epidural fat pad are varied at each disc level and CT findings of the normal disc and spinal canal areessential in diagnosis of herniation or bulging. However, there has been no previous publictions providing us withthe detailed criteria of normal disc and neural canal cross-sectional anatomy. The present study has been carriedout to investigate CT anatomy of the normal distribution of the epidural fat at each lumbar disc level to set up abaseline by which one will be able to judge disc herniation or bulging. The materials consisted of 30 adultvoluteers. They were sampled randomly from the individual undergoing a CT study for the reasons other than spinaldisease and having no symptoms or sings of the lumbar spinal disease. CT scans were performed with a SiemensSomatom 2 Scanner. Technical factors involved were: Tube voltage 125 KVp, exposure time 10 seconds, 460 mAs,256x256 matrix, and pixel size 0.4mm. Contiguous 2-mm thick slice of each disc were obtained parallel to the discplane at 3-mm interval for 5 lumbar discs. The results are follows; 1. The shape of the posterior disc margins atL1-2, L2-3, and L3-4 levels were concave anteriorly in over 90% of the cases compared with the disc margins atL4-5 and L5-S1 where they were convex or flat in over 85%. 2. None of the convex discs showed focal nodularprotrusion nor were more convex than the anterior margin of the disc. 3. The lower the lumbar levels, the greaterthe visibility of the epidural fat pad. The visibility of the epidural fat pad in both lateral reccesses was 81.5%at the L4-5 level and 100% at the L5-S1 level. There was no single case in which epidural fat pad was visible inonly one side. 4. The visibility of the epidural fat just behind the disc was only 18.5% at the L4-5 levelcompared with 88% at the L5-S1 level.